# How would a population react to an AI made for the purpose of statewide surveillance using current tech?

Setting

Scenario

In the near future, say 2030, with current technology, the Texas state government is implementing a unified surveillance program that will extensively use a variety methods, including :

• Street cameras
• Wi-fi location trackers
• 911 call data
• Stingray technology
• Criminal record/vehicle registration database
• Passive social media monitoring
• Web traffic interception (NSA tech)

in an attempt to give their state the advantage when it comes to gathering intelligence on the notorious cartels that are waging a territory war which has been ravaging the state. However, this program is different from others in one way : it utilizes an experimental AI developed jointly with SRI International, the DHS, MIT and the FBI. Named the ADAS (Analytic Data Assisted Surveillance), it has the power to track and identify suspicious behaviors or patterns by taking in data from all of the sources listed above.

Technical Details

This AI is nowhere near sentient. It is closer to the purpose built AIs that are strictly designed to function as an automated data analyst. Think of it as a less advanced version of the Superintendent from Halo ODST, connected to a surveillance grid that is a somewhat more intrusive version of modern day Camden's surveillance grid. It can detect and interpret complex patterns using pattern recognition, natural language processing, image processing and machine learning.

Examples of how it would reason :

import sqlite3

setloc(6700 Sherman Street)
cur = con.cursor()

scn_plate(E29346)
cur.execute(SELECT * FROM VEH_REGIST;)
df = pd.read_sql_query(cur.execute(WHERE plate_num = E29346;), con)

if theft_status == 1:
return(warn_veh_theft, plate_num, scan_loc)
veh_sus.append(plate_num)
else :
return()
veh_log.append(plate_num)

scn_face(ID9283498132475)
cur.execute(SELECT * FROM FED_PER_ID;)
df = pd.read_sql_query(cur.execute(WHERE id_no = ID9283498132475;), con)

if warrant_status == 'FELONY':
return(warn_felon, id_no, scan_loc)
act_felony.append(id_no)
elif warrant_status == 'MISDMR':
return(warn_misdmr, id_no, scan_loc)
act_misdmr.append(id_no)
else:
return()
per_log.append(id_no)


The Problem

How would this affect the lives of citizens? Each person knows that they are constantly being watched, so it would obviously have a chilling effect on dissenting speech, although that has already happened to some degree in my universe.

How would the cartels adapt? Assume that the larger ones have skilled hackers that are good enough to find holes in the system, and are skilled enough to keep up with the cat and mouse game of encryption against the government's programmers. Would they focus on disrupting the system, going back to old-school hand ciphers, or both?

What would the political ramifications be? Since whoever controls this system has a lot of power to do some serious muckraking, or cover it up as well...

This question asks for hard science. All answers to this question should be backed up by equations, empirical evidence, scientific papers, other citations, etc. Answers that do not satisfy this requirement might be removed. See the tag description for more information.

• upvote for including source code. – King-Ink Jan 26 '16 at 18:59
• Covering both the cartels and general population is a little broad. What about this communications filtering being used to monitor people based on their political views, not just criminal intent? Should that be taken into consideration as a likely scenario? – AndreiROM Jan 26 '16 at 19:08
• @AndreiROM Absolutely. I forgot to consider the political aspect, but it would definitely be a major factor as well. And it is overkill to monitor everyone, but that's part of the point. The drug war and paranoia has permeated so deep into our society that we can no longer tell who is a criminal, and who isn't. The divisions in society that are created stem far enough that the government can't even trust the average joe. – Carlos Danger Jan 26 '16 at 19:14
• One more question. You mention that the population in your world are already a bit more cowed into submission than us. It's difficult to estimate a response from the public if we don't know what might drive them over the edge. Would you be able to provide an example of a situation and just how far the people might bow to the demands of the government? Also, do they still have access to powerful fire-arms, like they currently do in the US? (especially in Texas) – AndreiROM Jan 26 '16 at 19:43
• Srriously? You're going to have a revolution on your hands. The US would slip into anarchy if these gangs had that much power. But it's not even realistic that these things would happen. – Xandar The Zenon Jan 26 '16 at 19:48

There will be a sort of bell curve response as people who don't have the skills or don't care settle into their day to day lives without worrying too much about intrusive surveillance, while the criminal, hacker and libertarian element are busy working out ways to circumvent it.

Some ideas are clearly "out". People hightailing it into the desert or other places where they think they can avoid surveillance will simply be telegraphic their intentions to the AI (and Predator drones can see you from a long way away). More subtle methods will have to be invented and practiced so that people can interact outside the parameters of the AI "in plain sight". Rather than smash and grabs, think more of con games, pickpockets and 3 card monte, or elaborate stage magic shows where the magician fools you with slight of hand, leaving you to wonder how exactly that was done.

Ubiquitous surveillance states are already in existence, and have had the better part of a century to practice. The former Soviet Union had things like FAX machines and photocopiers in locked rooms with limited access and hordes of people co-opted or forced to spy on their neighbours. East Germany had an incredible ratio of agents to population (over 174,000; over 2% of the population), and more modern states like Iran and China (and modern Russia) use the internet and social media monitoring and electronic warfare methodology to extensively monitor and filter the population's information.

Despite all this, there are a multitude of things people do, both high tech (various work arounds exist to penetrate the "Great Firewall of China", or you can be as simple was using "burn" phones for a single conversation), to low tech (Samizdat in the Soviet Union was basically conducted by hand copying information and passing sheets to trusted agents or through dead letter drops).

The real key is to ensure that your activities are not far enough outside the parameters considered "normal" to attract attention. If the backstory you wrote is accurate, there must have been a time where the various anti government groups infiltrated or subverted various government agencies, so can have inside people to monitor what the AI is doing and report back what is working and what isn't, or even manipulate the feeds or outputs of the AI itself.

• A combination of this and the two answers above would probably constitute an effective response to the system. Having programmers on the inside to change how the system works is a viable idea, but I also like the concept of eroding whatever trust is left in the program by manipulating the rather simple algorithms fool it into targeting innocent people. – Carlos Danger Jan 27 '16 at 18:39

People (and cartels) will probably begin to meet in place outside the sight of the AI and controls, like in countryside, a baseball game, a busy street, or a lake river and talk, while doing something completely legal and unrelated. Or they can use a method like the Italian Mafia, which used a code written on paper, the so called "pizzini".

In general, I suppose the the majority of the people probably will not care (not more than now) while the cartels will soon learn to avoid the surveillance, by technological countermeasures or going back to less technological methods that will probably beat the surveillance infrastructure

One possible cartel strategy might be to overload the system.

• Make a point to have not-yet-convicted kingpins use the same barbers, the same grocers, the same country clubs as respectable citizens and political leaders. Suddenly those people are flagged as associated with criminal circles.
• Say they know one of their operatives is under suspicion. They cut that operative loose from their real operations and send him out to hand-deliver advertising flyers to a dozen random and not-so-random people. "Good morning. Have you considered to change your phone provider? No? Sorry, have a nice day."
• In areas with suitable climate, build a drone which seeds/plants MJ in upper-class suburban gardens. "Sorry officer, no idea where that came from." "Yeah, sure."
• They take some of their not-yet-laundered money and wire it to random people. Couple of thousand dollars each, no comment.

All those people would become persons of interest to the AI, right? They show up on no-fly lists, based on classified intelligence, they are denied jobs in even remotely security-critical sectors, with no legal recourse. Until those citizens dismantle the surveillance operation.